CN102157155B - Representation method for multi-channel signal - Google Patents

Abstract

A multi-channel input signal having at least three original channels is represented by a parameter representation of the multi-channel signal. A first balance parameter (r1), a first coherence parameter, or a first inter-channel time difference between a first channel pair and a second bal-ance parameter (r2), or a second coherence parameter, or a second inter-channel time difference parameter between a second channel pair are calculated. This set of parameters is the parameter representation of the original signals. The first channel pair (102, 104) has two channels, which are different from two channels of a second channel pair (102, 104, 103). Furthermore, each channel of the two chan-nel pairs is one of the original channels, or a weighted combination of the original channels, and the first channel pair and the second channel pair include information on the three original channels. For multi-channel reconstruction purposes, the parameters are used in addition to down-mixing information to generate a selectable number of out-put channels in a scalable fashion.

Description

Multi-channel audio signal method for expressing

Technical field

The multichannel that the present invention relates to the sound signal to usage space parameter represents to encode.The present invention has lectured the new method for estimating and define proper parameter, from number is less than the sound channel of output channels number, reproduces multi-channel signal.Particularly, target of the present invention is to minimize the bit rate that multichannel represents, and provides the coded representation of the multi-channel signal that can realize data simple coding and decoding for all possible channel configuration.

Background technology

As PCT/SE02/01372 " Efficient and scalable Parametric Stereo Coding for Low Bitrate Audio Coding Applications " as shown in, can be from the given stereo image that the monophonic signal of stereo image compression expression, reproducing and original stereo image approach very much.Ultimate principle is that input signal is divided into frequency band and time period, to these frequency bands and time period, and degree of coherence (ICC) between intensity difference (IID) and sound channel between estimation sound channel.First parameter is the measurement on specific frequency band, power being distributed between two sound channels, and second parameter be on specific frequency band to two sound channels between the estimation of degree of coherence.In demoder one side, by according to IID-data, monophonic signal being distributed between two output channels and adding a de-correlated signals to retain the sound channel degree of coherence of original stereo sound channel, thereby stereo image is reproduced out from monophonic signal.

For the situation (multichannel refers to more than two output channels in context) of multichannel, have to consider some extra problems.There is at present several multichannel configuration.What know most is 5.1 configurations (middle sound channels, preposition left/right, around left/right, with LFE sound channel).But also there are many other configurations.According to the viewpoint of complete coder/decoder system, wish that the system obtaining can be used identical parameter set (for example IID and ICC) or its subset to all channel configuration.ITU-RBS.775 has defined mixed (down-mix) scheme of some contractings, can from given channel configuration, obtain the channel configuration by still less sound channel forms.Always have to all channel decodings and to the dependence difference that contracts mixed, wish that obtaining a kind of multichannel represents, makes receiver can extract the parameter relevant with existing channel configuration before to channel decoding.In addition, according to expanding or embedded encoded viewpoint, wishing to obtain inherent extendible parameter set, wherein, for example, is possible in the enhancement layer storage of the bit stream data corresponding with surround channel.

With above-mentioned contrary, what also hope obtained is the characteristic based on processed signal to adopt different parameter-definitions, between parametrization, switches, and this causes the minimal bit rate expense for current processed signal segment.

In this area, also have another kind of known employing to add with the multi-channel signal of signal or the mixed signal of contracting and additional parameter side-information and represent, be called two-channel prompting coding (BCC, binaural cue coding).Technique exists " Binaural Cue Coding-Partl:Psycho-Acoustic Fundamentals and Design Principles ＂, IEEE Transactions on Speech and Audio Processing, vol.11, No.6, November 2003, F.Baumgarte, C.Faller, with " Binaural Cue Coding.PartII:Schemes and Applications ＂, IEEE Transactions on Speech andAudio Processing vol.11, No.6, November2003, in C.FallerandF.Baumgarte, be described.

In general, two-channel prompting coding is a kind ofly to carry out according to a mixed audio track of contracting and side-information the method that multichannel space presents.Comprise amplitude difference between sound channel, relevant parameters between mistiming and sound channel between sound channel by the some parameters for audio frequency is rebuild and audio frequency presents that BCC scrambler calculates and BCC demoder uses.Between these sound channels, prompting is the deciding factor of aware space image.Multiple time samples pieces for original multi-channel signal provide these parameters, and providing of these parameters have frequency selectivity, make each sample block of multi-channel signal have several promptings to several frequency ranges.For the generalized case of C playback channels, sound channel between consider between sound channel the mistiming between amplitude difference and sound channel in each subband, that is to say with respect to consider each sound channel with reference to sound channel.Define a sound channel as with reference to sound channel for amplitude difference between each sound channel.By the mistiming between amplitude difference between sound channel and sound channel, a certain loudspeaker that source of sound is presented to used playback apparatus between any direction on be possible.Present range or the diffusion of source of sound in order to determine, concerning all audio tracks, each subband considers that a parameter is just enough.This parameter is exactly relevant parameters between sound channel.Present source of sound width control by revising subband signal, make all possible sound channel to thering is relevant parameters between identical sound channel.

In BCC coding, between all sound channels, amplitude difference is determined between with reference to sound channel 1 and other sound channel.For example, when central sound channel is confirmed as with reference to sound channel, by amplitude difference between the fourth sound road between amplitude difference and right surround channel and middle sound channel between the triple-track between amplitude difference, left surround channel and middle sound channel between the second sound channel between amplitude difference, R channel and middle sound channel between the first sound channel calculating between L channel and middle sound channel.This situation has been described the allocation plan of five-sound channel.Strengthen sound channel namely when known " supper bass " sound channel when five-sound channel allocation plan comprises a low frequency in addition, strengthen namely amplitude difference between the fifth sound road between unique reference sound channel of sound channel and middle sound channel by calculating low frequency.

When using a contracting mixing sound road to be also called as the sound channel of " monophony ", and while using transmission prompting as ICLD (amplitude difference between sound channel), ICTD (mistiming between sound channel) and ICC (being concerned with between sound channel) to rebuild original multichannel, by point out the spectral coefficient of amendment sound channel signal with these.Be used for determining the arithmetic number of amplitude correction for each spectral coefficient by use and complete amplitude correction.Be used for determining the plural numerical value of phase place correction for each spectral coefficient by use and generate the mistiming between sound channel.Another function has been determined degree of coherence impact.Each sound channel calculates by the factor of computing reference sound channel first for the factor of amplitude correction.Make each frequency part with reference to the calculating of the sound channel factor, the power sum of all sound channels equal signal and power.Then, according to the amplitude correction factor with reference to sound channel, the amplitude correction factor of other sound channels can with corresponding ICLD calculation of parameter out.

So, synthetic in order to complete BCC, calculate the amplitude correction factor with reference to sound channel.For this calculating, all ICLD parameters of a frequency band are all essential.Then,, according to the amplitude correction of this single sound channel, the namely non-amplitude correction factor with reference to sound channel of other sound channel will be calculated.

The shortcoming of this method is, in order to rebuild completely, needs amplitude difference between each sound channel.In the time there is the transmission channel of easily makeing mistakes, this requires to be just more a problem.Each mistake between the sound channel of transmission in amplitude difference will cause the mistake of the multi-channel signal of rebuilding, because need amplitude difference between each sound channel to calculate each sound channel of multichannel output signal.In addition, when between a sound channel, amplitude difference is lost in transmission, reconstruction is exactly impossible, although amplitude difference is only essential to for example left surround channel or right surround channel between this sound channel, and left around or right surround channel multichannel is rebuild very unimportant because most information has been included in left front put L channel that sound channel be known as afterwards, right front putting in the R channel or middle sound channel that sound channel be known as afterwards.When between the sound channel of low frequency reinforcement sound channel, amplitude difference is lost in transmission, and it is even even worse that this situation becomes.In this case, do not have or only vicious multichannel to rebuild be possible, not too have decisive although low frequency is strengthened sound channel to audience's sense of hearing comfort level.So the mistake between single sound channel in amplitude difference spreads into the mistake in each reconstruction output channels.

In addition, existing BCC scheme, at AES meeting article 5574, " Binaural Cue Coding applied to Stereo and Multi-channel Audio Compression ＂; C.Faller, F.Baumgarte, May10to13; 2002; Munich, the scheme of describing in Germany, when considering that intuition is while listening to scene; because only one with reference to sound channel, be not too suitable.It is factitious to the mankind that each things is related to single reference sound channel, and certainly, this is the final goal of whole audio frequency processing.Different, people has two ears, is positioned at the different both sides of head.So people's natural sense of hearing impression is, whether balance more left or more to the right of signal, or between front and back balance.With its by contrast, allow people feel in sound field certain sound source each loudspeaker with respect to single with reference to loudspeaker between whether in certain balance, this is factitious.When considering that bit rate requirements, extensibility requirement, requirement on flexibility, reconstruction manual signal require or when error robustness requirement, this one side is nature sense of hearing impression and be the negative results that difference between mathematics/physical model of BCC may cause encoding scheme on the other hand.

Summary of the invention

The object of this invention is to provide the improvement thought that represents multi-channel audio signal.

The present invention is based on following discovery, represent for multichannel, must rely on sound channel between balance parameters.And, also find by least two different balance parameters are provided, represent two different sound channels between the parameter of balance, the Parametric Representation of multi-channel signal is in the cards.Particularly, dirigibility, extensibility, error robustness are so that bit rate efficiency is all following true result, as first sound channel on the first balance parameters basis to second sound channel as the second balance parameters basis to being different, wherein, form four sound channels that these sound channels are right different.

So the thought of thought of the present invention and single reference sound channel is different, but adopted the thought of many balances or overbalance, this thought concerning people's auditory perception more intuitively, more natural.Particularly, to comprising original channel, contracting mixing sound road, or preferred as the sound channel on the first and second balance parameters bases, the particular combinations between input sound channel.

Have been found that, from as sound channel to sound channel the first sound channel and the original left sound channel to second sound channel and original right sound channel sum draw as sound channel balance parameters, for providing between middle sound channel, L channel and R channel, energy distribution is particularly useful accurately.Also will in context, note, these three sound channels have generally contained the most information of sound sight, and wherein particularly stereo location, left and right is not only subject to the impact of balance between left and right, is also subject to the impact of balance between middle sound channel and left and right acoustic channels sum.This is found by using this balance parameters based on the preferred embodiment of the present invention to reflect.

Preferably, in the time of a mixed signal of monophony contracting of transmission, also find, except in/left side adds right balance parameters, left/right balance parameters, rear left/rear right balance parameters and front/rear balance parameters are to have the best solution that efficient bit-rate parameters represents, this Parametric Representation is flexible, and anti-wrong ability is good, and has suitable fidelity.

At receiver side, in synthesizing with BCC, only calculate each sound channel by the information of transmission different, many balances of the present invention represent to utilize the information about the mixed scheme of contracting that is used for producing contracting mixing sound road.So according to the present invention, except balance parameters, the mixed scheme information of contracting not using in prior art systems is also used in upper mixed (up-mix).Therefore carry out upper mixed operation, the balance that makes to form in re-establishing multiple acoustic track signal between the sound channel that sound channel that balance parameters relies on is right is determined by this balance parameters.

This thought, namely forms different sound channels pair to different balance parameters, makes can obtain some sound channels the balance parameters of each transmission is not known to know in situation.Particularly, according to the present invention, in the situation that not knowing any rear left/rear right balance or front/rear balance, still can rebuild left and right and middle sound channel.This effect for the unusual extensibility of fine adjustments provide may because extract additional parameter or the additional balance parameters of transmission to receiver from bit stream, all will allow the one or more additional auditory channel of reconstruction.The single frame of reference of this and prior art is different, in the single frame of reference of prior art, needs amplitude difference between each sound channel to rebuild the whole or one group of sound channel in all reconstruction output channels.

Thought of the present invention also has dirigibility, and wherein the selection of balance parameters can adapt to specifically rebuild environment.For example, in the time that five-sound channel configuration structure forms original multi-channel signal structure, and in the time that quadraphonic structure forms re-establishing multiple acoustic track structure, wherein this structure only has a circulating loudspeaker, for example, after this loudspeaker is positioned at hearer, anterior-posterior balance parameter in the situation that left surround channel and right surround channel are not known for calculation combination surround channel provides possibility.This is with single different with reference to sound channel system, single with reference to must extract amplitude difference between the sound channel of amplitude difference and right surround channel between the sound channel of left surround channel in sound channel system from data stream.Then, must calculate left surround channel and right surround channel.Finally, two sound channels must be superposeed to obtain a circulating loudspeaker sound channel of quadraphonic reproduction structure.Above-mentioned in the balance parameters that has more intuitive sense and have more hearer's guiding represents is not to carry out in steps, because this expression is because balance parameters represents automatically to produce combination surround channel, this expression is not associated with single reference sound channel, but allows using the combination of original channel as the right sound channel of balance parameters sound channel.

The present invention relates to the problem that the parametrization multichannel of sound signal represents.Provide a kind of high efficiency mode to represent to define suitable parameter for multichannel, and provide a kind of needs to decode and just can extract the ability for representing the parameter of wanting sound channel all sound channels.The present invention has further solved and has selected Optimal parameter setting for given signal segment and make the spatial parameter of the given signal segment minimized problem of required bit rate of encoding.The present invention has also summarized and how to have applied before only to two of general multichannel environment applicable decorrelating methods of sound channel situations.

In a preferred embodiment, the present invention has following characteristics:

-in scrambler one side, multi-channel signal contracting is mixed to monophony or two-channel represents;

-under the condition of given multi-channel signal, definition represents the parameter of multi-channel signal, or flexibly take every frame as basis to minimize bit rate, or make demoder can on layer bitstreams, extract channel configuration;

-in demoder one side, the channel configuration that given demoder is current supported, extracts relevant parameter set;

-given current channel configuration, the signal of the mutual decorrelation of generation requirement;

-given the parameter set that decoding obtains from bitstream data and the signal of decorrelation, reproduce output signal.

The parameterized definition of-multi-channel audio signal, makes use and the sound channel configuration-independent of the subset of identical parameter or parameter.

The parameterized definition of-multi-channel audio signal can be applied parameter in scalable coding scheme, and wherein the subset of parameter set is can transmit in the different layers of extended data stream.

The parameterized definition of-multi-channel audio signal, makes the energy Reconstruction of decoder output signal, and the bottom audio codec that can not be used to the mixed signal that contracts to encode destroys.

Switching between-different multi-channel audio signal parametrizations, minimizes the bit-rate overhead for parametrization is encoded.

The parameterized definition of-multi-channel audio signal, wherein comprises and represents that contracting mixes the parameter of the energy correction factor of signal.

-reappear multi-channel signal with the decorrelator of some mutual decorrelations.

The upper mixed matrix H that the parameter set of-basis based on transmitted calculates, reappears multi-channel signal.

Accompanying drawing explanation

The present invention describes by illustrative example with reference to accompanying drawing, and scope of invention or spirit are not limited to this certainly, in accompanying drawing:

Fig. 1 illustrates the term for 5.1 channel configuration in the present invention;

Fig. 2 illustrates possible scrambler of the present invention and realizes;

Fig. 3 illustrates possible demoder of the present invention and realizes;

Fig. 4 illustrates of the present invention one preferred multi-channel signal Parameterization Scheme;

Fig. 5 illustrates of the present invention one preferred multi-channel signal Parameterization Scheme;

Fig. 6 illustrates of the present invention one preferred multi-channel signal Parameterization Scheme;

Fig. 7 illustrates the schematic structure of the mixed scheme of contracting that produces a basic sound channel or two basic sound channels;

Fig. 8 illustrates based on balance parameters of the present invention with about the schematic structure of the upper mixed scheme of the information of the mixed scheme of contracting;

Fig. 9 a illustrates determining of scrambler one side range parameter;

Fig. 9 b illustrates the use of demoder one side range parameter;

Figure 10 a illustrate have a parameterized different piece of multichannel at the different layers of bit stream can spread bit stream;

Figure 10 b illustrates extensibility form, and which sound channel indication can build with which balance parameters, and which balance parameters and sound channel are that do not use or unevaluated; And

Figure 11 illustrates the application based on mixed matrix of the present invention.

Embodiment

Following embodiment is only used to set forth the principle that the present invention represents about sound signal multichannel.To recognize, be apparent to the modifications and changes of layout described herein and details for others skilled in the art.Therefore, object is only the scope that is limited to the claim that is about to make, and is not limited to detail described herein and the explaination to embodiment.

In the following description of this invention, parametrization IID and ICC parameter how are summarized, and how represent with the multichannel that they reappear sound signal, wherein suppose that all signals that relate to are the subband signals in bank of filters, or in whole frequency ranges of corresponding sound channel, some other frequency selectivity of a part represent.So recognize, the invention is not restricted to specific bank of filters, and the frequency band that the present invention represents the subband of signal hereinafter summarizes, and identical operation is applicable to all subband signals.

Although balance parameters is known as " intensity difference between sound channel (IID) " parameter, be stressed that, sound channel between balance parameters must not be the ratio between energy or intensity and energy or the intensity of sound channel centering second sound channel of sound channel centering the first sound channel.Generally, balance parameters represents the location of sound source between two right sound channels of sound channel.Although this location is provided by energy/amplitude/intensity difference conventionally, also can use other characteristics of signal, as the power of two sound channels is measured or time or the frequency envelope etc. of sound channel.

The visible different sound channels for 5.1 channel configuration in Fig. 1, wherein a (t) 101 represents left surround channel, b (t) 102 represents the left front sound channel of putting, sound channel in c (t) 103, d (t) 104 represents the right front sound channel of putting, e (t) 105 represents right surround channel, and f (t) 106 represents LFE (low frequency sound effect) sound channel.

Suppose that definition expectation operator is:

$E\left[f\left(x\right)\right]=\frac{1}{T}\underset{0}{\overset{T}{\∫}}f\left(x\left(t\right)\right)\mathrm{dt}$

Can define according to following formula so the energy (here take left surround channel as example) of the sound channel of above-outlined:

A＝E[a ^２(t)]

These five sound channels scrambler one side contracted mixed (down-mix) represent or monophony represents to two sound channels.Can complete contracting by some methods mixed, normally used method is that ITU contracting is mixed, and it is defined as follows:

5.1 to two sound channel contractings are mixed:

l _d(t)＝αb(t)+βa(t)+γc(t)+δf(t)

r _d(t)＝αd(t)+βe(t)+γc(t)+δf(t)

5.1 is mixed to monophony contracting:

$m_d\left(t\right)=\sqrt{\frac{1}{2}}(l_d\left(t\right)+r_d\left(t\right))$

The accepted value of constant alpha, β, γ and δ is:

α=1， $\mathrm{\β}=\mathrm{\γ}=\sqrt{\frac{1}{2}}$ And δ=0.

IID parameter is defined as the energy ratio of two sound channels of selecting arbitrarily or sound channel set of weights.The channel energies of 5.1 channel configuration of given above-outlined, can define some groups of IID parameters.

Fig. 7 has shown that general contracting mixes device 700, and it has used above-mentioned formula to ask and has calculated single sound channel m or two preferably based on stereosonic sound channel 1 _dand r _d.General, the mixed device of contracting uses the mixed information of specific contracting.In the mixed embodiment of preferred linear contracting, the mixed information of contracting comprises weighted factor, β, γ and δ.Be known in the art, can use more or less constant or the weighting factor of non-constant.

During the contracting of recommending at ITU mixes, α is made as 1, β and γ is made as equal value, be all 0.5 square root, and δ is made as 0.In general, the α factor can change between 1.5 to 0.5.In addition, β and γ can be unequal mutually, and change between 0 and 1.It is the same that low frequency is strengthened to sound channel f (t).The factor delta of this sound channel can change between 0 and 1.In addition, do not need to be equal to each other for the mixed mixed factor of right contracting of left contracting.When considering that non-automatic contracting is mixed, for example, while execution by sound slip-stick artist, it is very clear that this situation becomes.It is mixed that sound slip-stick artist more focuses on carrying out the contracting of novelty, rather than the contracting of being arranged by Mathematical rule is mixed.Different, sound slip-stick artist is subject to the domination of his novelty impression.Mixed by after specifically parameter set records when this " novelty " contracting, it will be used by upper mixed (up-mix) device of invention according to the present invention, and as shown in Figure 8, on this, mixed device is arranged by parameter not only, and the extraneous information of the mixed scheme that also contracted is arranged.

When as shown in Figure 7, carry out linear contracting mixed, weighting parameters is exactly the preference information about the mixed scheme of contracting, will be used by upper mixed device.But in the time being used in out of Memory in the mixed scheme of contracting and existing, these information also can be used as about the information of the mixed scheme of contracting and are used by upper mixed device.Such out of Memory, for example, can be specificity factor or the function in particular matrix element or the matrix element in upper mixed matrix, for example, and as shown in Figure 11.

5.1 channel configuration of general introduction in Given Graph 1, and notice that how associated with 5.1 channel configuration other channel configuration is: for the situation of triple-track, wherein surround channel cannot obtain, and that is to say, according to above-mentioned symbol, B, C and D can obtain.For quadraphonic configuration, B, C and D can obtain, and represent that the A of single surround channel and the combination of E also can obtain, and it is more frequent is in this article represented as rearmounted sound channel.

The present invention has defined the IID parameter that is applicable to all these sound channels, that is to say, the quadraphonic subset of 5.1 channel configuration has corresponding subset in the IID parameter set of describing 5.1 sound channels.

Following IID parameter set has solved this problem:

${\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=\frac{L}{R}=\frac{{\mathrm{\&alpha;}}^{2}B+{\mathrm{\&beta;}}^{2}A+{\mathrm{\&gamma;}}^{2}C+{\mathrm{\&delta;}}^{2}F}{{\mathrm{\&alpha;}}^{2}D+{\mathrm{\&beta;}}^{2}E+{\mathrm{\&gamma;}}^{2}C+{\mathrm{\&delta;}}^{2}F}$

${\mathrm{<figure-callout\; id="2"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_2=\frac{{\mathrm{\&gamma;}}^{2}2C}{{\mathrm{\&alpha;}}^2(B+D)}$

${\mathrm{<figure-callout\; id="3"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_3=\frac{{\mathrm{\&beta;}}^2(A+E)}{{\mathrm{\&alpha;}}^2(B+D)+{\mathrm{\&gamma;}}^{2}2C}$

${\mathrm{<figure-callout\; id="4"\; label="r"\; filenames="HSA00000475998500091.png"\; state="\{\{state\}\}">r</figure-callout>}}_4=\frac{{\mathrm{\&beta;}}^{2}A}{{\mathrm{\&beta;}}^{2}E}=\frac{A}{E}$

${\mathrm{<figure-callout\; id="5"\; label="r"\; filenames="HSA00000475998500041.png,HSA00000475998500082.png"\; state="\{\{state\}\}">r</figure-callout>}}_5=\frac{{\mathrm{\&delta;}}^{2}2F}{{\mathrm{\&alpha;}}^2(B+D)+{\mathrm{\&beta;}}^2(A+E)+{\mathrm{\&gamma;}}^{2}2C}$

Obviously, r ₁energy Ratios between the corresponding left contracting mixing sound road of parameter and right contracting mixing sound road.R ₂energy Ratios in parameter correspondence between the preposition sound channel in sound channel and left and right.R ₃energy Ratios between corresponding three the preposition sound channels of parameter and two surround channels.R ₄energy Ratios between corresponding two surround channels of parameter.R ₅parameter correspondence the energy Ratios between LFE sound channel and all other sound channels.

Fig. 4 illustrates the energy Ratios of explaining above.Different output channels is by 101 to 105 expressions, the same with in Fig. 1, is therefore not described in detail here.Loudspeaker setting is divided into the left half and the right half part, and wherein middle sound channel 103 respectively belongs to the part in two halves.According to the present invention, the energy Ratios between left demifacet and right-half plane is exactly exactly the parameter r mentioning ₁.This is by r in Fig. 4 ₁solid line below represents.In addition, according to the present invention, in middle sound channel 103, left frontly put 102 and right frontly put energy distribution between 104 by r ₂represent.Finally, the energy distribution between whole preposition sound channel setting (102,103 and 104) and rearmounted sound channels (101 and 105) by Fig. 4 in r ₃shown in the arrow on parameter side.

The energy in the single contracting mixing sound road of given above Parameterization Scheme and transmission:

$M=\frac{1}{2}({\mathrm{\&alpha;}}^2(B+D)+{\mathrm{\&beta;}}^2(A+E)+{2\mathrm{\&gamma;}}^{2}C+{2\mathrm{\&delta;}}^{2}F),$

The energy of reconstructed channels can be expressed as:

$F=\frac{1}{2{\mathrm{\&gamma;}}^2}\frac{{\mathrm{<figure-callout\; id="5"\; label="r"\; filenames="HSA00000475998500041.png,HSA00000475998500082.png"\; state="\{\{state\}\}">r</figure-callout>}}_5}{1+{\mathrm{<figure-callout\; id="5"\; label="r"\; filenames="HSA00000475998500041.png,HSA00000475998500082.png"\; state="\{\{state\}\}">r</figure-callout>}}_5}2M$

$A=\frac{1}{{\mathrm{\&beta;}}^2}\frac{{\mathrm{<figure-callout\; id="4"\; label="r"\; filenames="HSA00000475998500091.png"\; state="\{\{state\}\}">r</figure-callout>}}_4}{1+r_4}\frac{{\mathrm{<figure-callout\; id="3"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_3}{1+{\mathrm{<figure-callout\; id="3"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_3}\frac{1}{1+{\mathrm{<figure-callout\; id="5"\; label="r"\; filenames="HSA00000475998500041.png,HSA00000475998500082.png"\; state="\{\{state\}\}">r</figure-callout>}}_5}2M$

$E=\frac{1}{{\mathrm{\&beta;}}^2}\frac{1}{1+r_4}\frac{{\mathrm{<figure-callout\; id="3"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_3}{1+{\mathrm{<figure-callout\; id="3"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_3}\frac{1}{1+{\mathrm{<figure-callout\; id="5"\; label="r"\; filenames="HSA00000475998500041.png,HSA00000475998500082.png"\; state="\{\{state\}\}">r</figure-callout>}}_5}2M$

$C=\frac{1}{{2\mathrm{\&gamma;}}^2}\frac{{\mathrm{<figure-callout\; id="2"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_2}{1+{\mathrm{<figure-callout\; id="2"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_2}\frac{1}{1+{\mathrm{<figure-callout\; id="3"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_3}\frac{1}{1+{\mathrm{<figure-callout\; id="5"\; label="r"\; filenames="HSA00000475998500041.png,HSA00000475998500082.png"\; state="\{\{state\}\}">r</figure-callout>}}_5}2M$

$B=\frac{1}{{\mathrm{\&alpha;}}^2}(2\frac{{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">r</figure-callout>}}_1}{1+r_1}M-{\mathrm{\&beta;}}^{2}A-{\mathrm{\&gamma;}}^{2}C-{\mathrm{\&delta;}}^{2}F)$

$D=\frac{1}{{\mathrm{\&alpha;}}^2}(2\frac{1}{1+r_1}M-{\mathrm{\&beta;}}^{2}E-{\mathrm{\&gamma;}}^{2}C-{\mathrm{\&delta;}}^{2}F)$

So the energy of M signal can be distributed in reconstructed channels, cause reconstructed channels to there is the energy identical with original channel.

In Fig. 8, illustrate above-mentioned preferred mixed scheme.According to the formula about F, A, E, C, B and D, obviously, the information about the mixed scheme of contracting that upper mixed device uses is exactly weighted factor, β, γ and δ, these factors are used for original channel to be weighted, and then weighting like this or unweighted sound channel are superimposed or are subtracted each other each other, to obtain the contracting mixing sound road of some, the quantity in these contracting mixing sound roads is less than the quantity of original channel.So known according to Fig. 8, according to the present invention, the energy of reconstructed channels is not only determined by the balance parameters that is transferred to demoder one side from scrambler one side, also mixes factor-alpha by contracting, beta, gamma and δ determine.

In the time considering Fig. 8, apparent, for calculating left and right channel energies B and D, in formula, use channel energies F, the A, E, the C that have calculated.But this also not necessarily means the upper mixed scheme of order.On the contrary, for obtaining the upper mixed scheme of complete parallel, the scheme that the specific mixed matrix that for example contains specific mixed matrix element by utilization is carried out, the formula of A, C, E and F is inserted in the formula of B and D.So apparent, reconstructed channels energy is only for example contracted and is mixed the factor and determine by balance parameters, contracting mixing sound road and the information about the mixed scheme of contracting.

Given above IID parameter, it is evident that, definition is resolved for the problem of the parameter set of the IID parameter of some channel configuration, and this shows clearly hereinafter.As an example, note triple-track configuration (that is to say, reappear three preposition sound channels from an available sound channel), it is evident that r ₃, r ₄and r ₅parameter is useless, because A, E and F sound channel do not exist.Also clearly, parameter r ₁and r ₂enough for reappear triple-track from the mixed single sound channel that contracts, because r ₁what describe is the energy Ratios between the preposition sound channel in left and right, and r ₂what describe is the energy Ratios between the preposition sound channel of middle sound channel and left and right.

In a more general case, be very easy to find IID parameter (r defined above ₁r ₅) be applicable to reappear all subsets of n sound channel from m sound channel.Wherein m < n≤6.Note Fig. 4, can think:

-for the system of reappearing 2 sound channels from 1 sound channel, from r ₁in parameter, obtain sufficient information, keep energy Ratios correct between sound channel.

-for the system of reappearing 3 sound channels from 1 sound channel, from r ₁and r ₂in parameter, obtain sufficient information, keep energy Ratios correct between sound channel.

-for the system of reappearing 4 sound channels from 1 sound channel, from r ₁, r ₂and r ₃in parameter, obtain sufficient information, keep energy Ratios correct between sound channel.

-for the system of reappearing 5 sound channels from 1 sound channel, from r ₁, r ₂, r ₃and r ₄in parameter, obtain sufficient information, keep energy Ratios correct between sound channel.

-for the system of reappearing 5.1 sound channels from 1 sound channel, from r ₁, r ₂, r ₃, r ₄and r ₅in parameter, obtain sufficient information, keep energy Ratios correct between sound channel.

-for the system of reappearing 5.1 sound channels from 2 sound channels, from r ₂, r ₃, r ₄and r ₅in parameter, obtain sufficient information, keep energy Ratios correct between sound channel.

Table in Figure 10 b illustrates above-mentioned extensibility feature.In Figure 10 a, illustrate and will can spread bit stream also go for the table in Figure 10 b what explain after a while, obtaining than accurate many extensibilities shown in Figure 10 a.

Thought of the present invention has special advantage for following situation, other balance parameters is not known also and is not extracted, only from single balance parameters r ₁in just can easily rebuild left and right sound channel.For this reason, in Fig. 8, in the formula of B and D, sound channel A, C, E and F are simply made as 0.

As selection, when only considering balance parameters r ₂time, reconstructed channels is the stack between middle sound channel and low-frequency channel (in the time that this sound channel is not made as 0) on the one hand, is the stack between the sound channel of left and right on the other hand.So only use single parameter, sound channel in can rebuilding on the one hand, can rebuild monophonic signal on the other hand.This feature can represent to produce use to simple triple-track, and wherein left signal and right signal obtain from the stack of left and right acoustic channels, for example binary, and energy between middle sound channel and left and right acoustic channels sum is by balance parameters r ₂accurately determine.

In context, balance parameters r ₁or r ₂be positioned at low expansion (scaling) layer.

About second entry showing in Figure 10 b, represent 3 sound channel B, D and C and F's and how to adopt two balance parameters rather than 5 balance parameters to obtain, these parameters r ₁and r ₂one of may be in the parameter r than being arranged in lower extension layer ₁or r ₂in high extension layer.

When consider in Fig. 8 formula time, apparent, in order to calculate C, the parameter r not extracting ₅with other parameter r not extracting ₃be set as 0.In addition, untapped sound channel A, E, F are also set as 0, make to calculate three sound channel B, D and middle sound channel C and low frequency strengthen sound channel F's and.

When the expression of 4 sound channels when mixed, is only extracted parameter r by upper from supplemental characteristic stream ₁, r ₂and r ₃just enough.In context, r ₃can be than other parameter r ₁or r ₂in next higher extension layer.It is specially suitable that 4 channel configuration represent to combine with super balance parameters of the present invention because as below by the description of carrying out together with Fig. 6, the 3rd balance parameters r ₃from the combination of the preposition sound channel on one side and the rearmounted sound channel of another side, obtain.This is because such fact, i.e. parameter r ₃anterior-posterior balance parameters, it be according to the combination with rearmounted sound channel A and E as the combination of the first sound channel and L channel B, R channel E and middle sound channel C as the sound channel of preposition sound channel to obtaining.

So the combined channels energy of two surround channels does not have in addition independent calculating and combination subsequently just can automatically obtain, and single with reference to channel structure in, the just situation of the other independent calculating of existence and combination subsequently.

In the time must reappearing 5 sound channel from single sound channel, other balance parameters r ₄essential.Parameter r ₄can be again in the higher extension layer of the next one.

In the time must carrying out 5.1 reconstruction, each balance parameters needs.Therefore, comprise next balance parameters r ₅the higher extension layer of the next one receiver of having to be sent to, and received machine estimation.

But, make to use the same method, expand IID parameter according to the expansion number of sound channel, above-mentioned IID parameter can be by expansion, covers channel number than 5.1 also large channel configuration s of the channel number that configure.Therefore, the invention is not restricted to above-mentioned example.

The situation of attention is now, and channel configuration is 5.1 channel configuration, and this is the most frequently used situation.And, suppose that 5.1 sound channels rebuild from two sound channels.In this case, can be by following parameter replacement parameter r ₃and r ₄define different parameter sets:

${\mathrm{<figure-callout\; id="3"\; label="q"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">q</figure-callout>}}_3=\frac{{\mathrm{\&beta;}}^{2}A}{{\mathrm{\&alpha;}}^{2}B}$

${\mathrm{<figure-callout\; id="4"\; label="q"\; filenames="HSA00000475998500091.png"\; state="\{\{state\}\}">q</figure-callout>}}_4=\frac{{\mathrm{\&beta;}}^{2}E}{{\mathrm{\&alpha;}}^{2}D}$

Parameter q ₃and q ₄represent respectively the energy Ratios between preposition and rearmounted L channel, and energy Ratios between preposition and rearmounted R channel.It is contemplated that some other Parameterization Scheme.

The visible Parameterization Scheme of revising in Fig. 5.Not to adopt energy distribution between the preposition and rearmounted sound channel of Summary of Parameters (in Fig. 4 by r ₃general introduction) and parametric description left side around and right surround channel between energy distribution (as in Fig. 4 by r ₄general introduction), operation parameter q ₃and q ₄describe and left frontly put 102 and left around the energy Ratios between 101 sound channels, and right frontly put 104 and right around the energy Ratios between 105 sound channels.

The present invention has lectured and can represent multi-channel signal with some parameter sets.Another feature of the present invention is that different parametrizations can rely on the quantification type of operation parameter to select.

For example, the system due to high bit rate constraints to parameter rudenss quantization, must make upper sneak out journey in the wrong parametrization not being exaggerated.

Note above-mentionedly reappearing the system of 5.1 sound channels two for rebuilding the expression of energy from a sound channel:

$B=\frac{1}{{\mathrm{\&alpha;}}^2}(2\frac{{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">r</figure-callout>}}_1}{1+r_1}M-{\mathrm{\&beta;}}^{2}A-{\mathrm{\&gamma;}}^{2}C-{\mathrm{\&delta;}}^{2}F)$

$D=\frac{1}{{\mathrm{\&alpha;}}^2}(2\frac{1}{1+r_1}M-{\mathrm{\&beta;}}^{2}E-{\mathrm{\&gamma;}}^{2}C-{\mathrm{\&delta;}}^{2}F)$

Obviously, due to the very little quantification effect of M, A, C and F parameter, subtraction operation will produce the great variety of B and D energy.

According to the present invention, should use different parametrizations, it is more insensitive to the quantification of parameter.So, if adopt rough quantification, parameter r ₁definition above:

${\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=\frac{L}{R}=\frac{{\mathrm{\&alpha;}}^{2}B+{\mathrm{\&beta;}}^{2}A+{\mathrm{\&gamma;}}^{2}C+{\mathrm{\&delta;}}^{2}F}{{\mathrm{\&alpha;}}^{2}D+{\mathrm{\&beta;}}^{2}E+{\mathrm{\&gamma;}}^{2}C+{\mathrm{\&delta;}}^{2}F}$

Can replace with another definition according to following formula:

${\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=\frac{B}{D}$

Produced the formula of rebuilding energy according to following formula:

$B=\frac{1}{{\mathrm{\&alpha;}}^2}\frac{{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">r</figure-callout>}}_1}{1+{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">r</figure-callout>}}_1}\frac{1}{1+{\mathrm{<figure-callout\; id="2"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_2}\frac{1}{1+{\mathrm{<figure-callout\; id="3"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_3}\frac{1}{1+{\mathrm{<figure-callout\; id="5"\; label="r"\; filenames="HSA00000475998500041.png,HSA00000475998500082.png"\; state="\{\{state\}\}">r</figure-callout>}}_5}2M$

$D=\frac{1}{{\mathrm{\&alpha;}}^2}\frac{1}{1+{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">r</figure-callout>}}_1}\frac{1}{1+{\mathrm{<figure-callout\; id="2"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_2}\frac{1}{1+{\mathrm{<figure-callout\; id="3"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_3}\frac{1}{1+{\mathrm{<figure-callout\; id="5"\; label="r"\; filenames="HSA00000475998500041.png,HSA00000475998500082.png"\; state="\{\{state\}\}">r</figure-callout>}}_5}2M$

And the formula of reconstruction A, E, C and F energy and identical above.Obviously,, from the angle quantizing, this parametrization has represented the better system of situation.

In Fig. 6, illustrate the energy Ratios of setting forth above.Different output channels is by 101 to 105 expressions, and identical with Fig. 1, so be not described in further detail here.Loadspeaker structure is divided into preposition part and rearmounted part.Energy distribution between whole preposition channel structure (102,103 and 104) and rearmounted sound channel (101 and 105), by parameter r in Fig. 6 ₃the arrow representing illustrates.

Another noticeable important feature of the present invention is when noting following parametrization

${\mathrm{<figure-callout\; id="2"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_2=\frac{{\mathrm{\&gamma;}}^{2}2C}{{\mathrm{\&alpha;}}^2(B+D)}$

${\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=\frac{B}{D}$

This is not only to see the better system of situation from the angle of quantification.Above-mentioned parametrization also has such advantage, can under the condition that does not affect surround channel, obtain for the parameter of rebuilding three preposition sound channels.It is contemplated that parameter r ₂relation between sound channel and all other sound channels is described.But shortcoming is to have comprised surround channel in to the estimation of the parameter of describing preposition sound channel.

Remember, the parametrization of describing in the present invention can be used for measuring the degree of correlation or the degree of coherence between sound channel, apparent, to r ₂calculating in comprise rearmounted sound channel, by the great negative effect of successful generation to the preposition sound channel of accurate reproduction.

For example, can imagine in all preposition sound channels signal identical and in rearmounted sound channel the complete incoherent situation of signal.Be usually used in reappearing the environmental information of original sound for rearmounted sound channel, this situation is unrare.

If compared with all other sound channels, the degree of correlation measured value between middle sound channel and all other sound channel sums is just very low, because rearmounted sound channel is completely uncorrelated by middle sound channel.Also be same for the parameter of estimating the degree of correlation between preposition left/right sound channel and rearmounted left/right sound channel.

So, the reconstruction energy that the parametrization reaching can be correct, but not comprising all preposition sound channels identical is the information of strong correlation.The information that it comprises is, the preposition sound channel in left and right and rearmounted sound channel decorrelation, and middle sound channel also with rearmounted sound channel decorrelation.But, from this parametrization, can not release all identical facts of preposition sound channel.

Overcoming of this problem can be by being used following parametrization

${\mathrm{<figure-callout\; id="2"\; label="r"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">r</figure-callout>}}_2=\frac{{\mathrm{\&gamma;}}^{2}2C}{{\mathrm{\&alpha;}}^2(B+D)}$

${\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=\frac{B}{D}$

As the present invention lectures, because demoder one side is used for to the estimation of the parameter of reappearing preposition sound channel, do not comprise rearmounted sound channel.

According to the present invention, in middle sound channel 103 and left frontly put sound channel 102 and right frontly put energy distribution between sound channel 104 by r ₂represent.Energy distribution between left surround channel 101 and right surround channel 105 is by r ₄explaination.Finally, put sound channel 102 and right frontly put energy distribution between sound channel 104 by r left front ₁provide.Obviously, except r1, all parameters are same as shown in Figure 4, herein r ₁corresponding is left frontly puts loudspeaker and right front energy distribution of putting between loudspeaker, but not energy distribution between whole left side and whole right side.For integrality, provide parameter r ₅the distribution of energy between middle sound channel 103 and L channel 106 described.

Fig. 6 has provided the summary of the preferred parametrization embodiment of the present invention.The first balance parameters r ₁(being represented by solid line) forms front left/front right balance parameters.The second balance parameters r ₂for in/left and right balance parameters.The 3rd balance parameters r ₃form front/rear balance parameters.The 4th balance parameters r ₄form rear left/rear right balance parameters.Last the 5th balance parameters r ₅in formation/LFE balance parameters.

Fig. 4 shows relevant situation.The the first balance parameters r explaining with solid line in the situation that contracting mixing left/right balance in Fig. 4 ₁, can replace with the original front left/front right balance parameters defining between as the right sound channel B of bottom sound channel and D.This parameter is by the dotted line r in Fig. 4 ₁diagram, corresponding solid line r in Fig. 5 and Fig. 6 ₁.

In two basic sound channel situations, parameter r ₃and r ₄, namely front/rear balance parameters and rear left/right balance parameters replace with two one-sided front/rear parameters.The first one-sided front/rear parameter q ₃also can be considered to the first balance parameters, it draws from the sound channel centering being made up of left surround channel A and L channel B.Second one-sided before/left balance parameters is parameter q ₄, it can be considered to the second parameter, and it is the second sound channel pair based on being made up of R channel D and right surround channel E.In addition, two sound channels are to separate.Centering/left and right balance parameters r ₂also be same, wherein, middle sound channel C is as the first sound channel, left and right sound channel B and D's and as second sound channel.

Another kind is applicable to rudenss quantization and is used for the parametric method of from one or two sound channel system of playback system 5.1 sound channels, is defined as follows according to the present invention.

For 1 to 5.1 sound channel:

${\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">q</figure-callout>}}_1=\frac{{\mathrm{\&beta;}}^{2}A}{M},{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">q</figure-callout>}}_2=\frac{{\mathrm{\&alpha;}}^{2}B}{M},{\mathrm{<figure-callout\; id="3"\; label="q"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">q</figure-callout>}}_3=\frac{{\mathrm{\&gamma;}}^{2}C}{M},{\mathrm{<figure-callout\; id="4"\; label="q"\; filenames="HSA00000475998500091.png"\; state="\{\{state\}\}">q</figure-callout>}}_4=\frac{{\mathrm{\&alpha;}}^{2}D}{M},{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">q</figure-callout>}}_2=\frac{{\mathrm{\&beta;}}^{2}E}{M}$ And ${\mathrm{<figure-callout\; id="5"\; label="q"\; filenames="HSA00000475998500041.png,HSA00000475998500082.png"\; state="\{\{state\}\}">q</figure-callout>}}_5=\frac{{\mathrm{\&delta;}}^{2}F}{M}$

Situation for 2 to 5.1 sound channels:

${\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">q</figure-callout>}}_1=\frac{{\mathrm{\&beta;}}^{2}A}{L},{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">q</figure-callout>}}_2=\frac{{\mathrm{\&alpha;}}^{2}B}{L},{\mathrm{<figure-callout\; id="3"\; label="q"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">q</figure-callout>}}_3=\frac{{\mathrm{\&gamma;}}^{2}C}{M},{\mathrm{<figure-callout\; id="4"\; label="q"\; filenames="HSA00000475998500091.png"\; state="\{\{state\}\}">q</figure-callout>}}_4=\frac{{\mathrm{\&alpha;}}^{2}D}{R},{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">q</figure-callout>}}_2=\frac{{\mathrm{\&beta;}}^{2}E}{R}$ And ${\mathrm{<figure-callout\; id="5"\; label="q"\; filenames="HSA00000475998500041.png,HSA00000475998500082.png"\; state="\{\{state\}\}">q</figure-callout>}}_5=\frac{{\mathrm{\&delta;}}^{2}F}{M}$

Obviously, the parameter that parametrization above comprises is more than the energy of signal transmission correctly being redistributed to the needed parameter of reproducing signal from strict point of theory.But parametrization is very insensitive to quantization error.

The above-cited parameter set for two basic channel structure, has used some with reference to sound channel.But different from the parameter configuration in Fig. 6, in Fig. 7, parameter set only relies on contracting mixing sound road rather than original channel as with reference to sound channel.Balance parameters q ₁, q ₃and q ₄draw from diverse parameter centering.

Although described the embodiment of some inventions, wherein for the sound channel that draws balance parameters to only comprising original channel (Fig. 4, Fig. 5, Fig. 6) or comprise original channel and contracting mixing sound road (Fig. 4, Fig. 5) or as shown in Fig. 7 bottom, only rely on contracting mixing sound road as with reference to sound channel, but preferably, in Fig. 2, can operate around parameter generators included in data encoder 206 combination that only uses original channel or original channel, and Non-equilibrium parameter based on the basic passage of passage centering passage or the combination of basic passage.This be because in fact can not guarantee completely from around scrambler to the transmission of surround decoder device, there is not the variation of energy in single basic sound channel or two stereo basic sound channels.Be operated under the condition of low bit rate, audio coder 205 (Fig. 2) or audio decoder 302 (Fig. 3) can cause the energy variation in multiple contracting mixing sound roads or single contracting mixing sound road.Such situation can cause the energy process of monophony contracting mixing sound road or stereo downmix sound channel, and the processing between the stereo downmix sound channel of left and right can be not identical, or can be even frequency selectivity or time selectivity.

According to the present invention, completely reliable in order to make for these energy variation, be each or each frequency band in each contracting mixing sound road, transmit additional range parameter.In the time that balance parameters is based on original signal rather than contracts mixed signal, single. a modifying factor says enough to each bringing, because any energy correction can not have influence on the equilibrium condition between original channel.Even in the time not transmitting additional range parameter, any contracting mixing sound road energy variation can not cause the location of mistake of source of sound in AV, and only can cause general loudness to change, loudness changes irritating unlike the movement of the source of sound being caused by the equilibrium condition changing.

Importantly, note carefully making energy M (contracting mixing sound road) is energy B, D, A, E, C and F sum, as mentioned above.Not such situation always, because mixed the phase dependent between the different sound channels of a sound channel by contracting.The energy correction factor can be used as additional parameter r _mtransmission, and the energy of the mixed signal of the contracting of receiving at demoder one side joint is defined as:

$r_{M}M=\frac{1}{2}({\mathrm{\&alpha;}}^2(B+D)+{\mathrm{\&beta;}}^2(A+E)+{2\mathrm{\&gamma;}}^{2}C+2{\mathrm{\&delta;}}^{2}F)$

In Fig. 9, summarize additional parameter r _mapplication.Contracted mixed input signal before sending it to the upper mixed module of 701-705, first by r in 901 _mparameter modification.Identical in this and Fig. 7, so no longer describe in further detail.It will be apparent to those skilled in the art that the parameter r for the mixed example of above-mentioned monophony contracting _m, can expand to one, road of each contracting mixing sound parameter, and therefore also be not limited to single contracting mixing sound road.

Fig. 9 a illustrates the range parameter counter 900 of invention, and Fig. 9 b has shown the amplitude modifier 902 of invention.Fig. 9 a has shown the situation of scrambler one side, and Fig. 9 b has shown the corresponding situation of demoder one side.Range parameter or " adding " parameter r _mfor modifying factor, provide specific energy Ratios.In order to explain above content, suppose following typical scene.To specific original multi-channel signal, there is on the one hand " main contracting is mixed ", there is on the other hand " parameter contracting is mixed ".Main contracting is mixed to be generated based on for example subjective tonequality impression in sound equipment workplace by Sound Engineer.In addition, specific audio storage medium also comprises that parameter contracting is mixed, mixed the carrying out around scrambler 203 by for example Fig. 2 of parameter contracting.A mixed basic sound channel or two the basic sound channels of comprising of parameter contracting, basic sound channel forms the basis that multichannel is rebuild, and multichannel is rebuild and has been used the balance parameters collection of original multi-channel signal or other Parametric Representation.

Can be such situation, for example, from transmitted from transmitter to receiver, broadcasting equipment think that transformation parameter contracting is mixed but it is mixed to transmit main contracting yet.In addition, mixed be upgraded to multichannel and represent for offices and men contract, broadcasting equipment also transmits the Parametric Representation of original multi-channel signal.Because energy (in a frequency band or in a piece) can (and typically will) change between the mixed parameter contracting of main contracting is mixed, relative amplitude parameter r _min square 900, produce, and be transferred to receiver as additional parameter.Range parameter is according to mixed the drawing of the mixed parameter contracting of main contracting, and the ratio of the energy in a mixed piece and the frequency band of the mixed parameter contracting of main contracting preferably.

In general, range parameter is calculated as energy and the (E of original channel _orig) mix the ratio of channel energies with contracting, wherein contracting mixing sound road can be the mixed (E of parameter contracting _pD) or the mixed (E of main contracting _mD) or the mixed signal of other contracting.Typically, use the mixed signal energy of specific contracting, it is transferred to demoder from scrambler.

Fig. 9 b illustrates the realization that demoder one side range parameter uses.The mixed signal of range parameter and contracting is input to amplitude modifier piece 902.Amplitude modifier, according to range parameter, is revised single basic sound channel or many basic sound channels.Because additional parameter r _mbe a relative value, this relative value has been multiplied by the energy of corresponding basic sound channel.

Although in the situation that Fig. 9 a and 9b show, amplitude correction is applied to contracting mixing sound road, range parameter also can be integrated in mixed matrix.For this reason, in Fig. 8 equation, each appearance of M is all replaced by " a r _mm ".

Research reappears the situation of 5.1 sound channels from two sound channels, is found as follows.

If the present invention uses together with the 302 bottom audio codecs of summarizing with 205 of Fig. 2 and Fig. 3, need to make more consideration.Note the previously IID parameter of definition, wherein r ₁definition basis

${\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=\frac{L}{R}=\frac{{\mathrm{\&alpha;}}^{2}B+{\mathrm{\&beta;}}^{2}A+{\mathrm{\&gamma;}}^{2}C+{\mathrm{\&delta;}}^{2}F}{{\mathrm{\&alpha;}}^{2}D+{\mathrm{\&beta;}}^{2}E+{\mathrm{\&gamma;}}^{2}C+{\mathrm{\&delta;}}^{2}F}$

This parameter demoder one side can be recessive obtain because system is reappeared 5.1 sound channels from two sound channels, if two transmission sound channels stereo downmix that are surround channel.

But the audio codec of working under certain bit rate constraints can be revised spectrum distribution, the L that demoder one side is recorded is different in the value of scrambler one side from them with R energy.According to the present invention, this impact on energy distribution in reproduction sound channel is eliminated by transmitting following parameter:

${\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HSA00000475998500061.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=\frac{B}{D}$

When rebuild 5.1 sound channel from two sound channels time, it is also such situation.

If signaling (signaling) device is provided, scrambler is by using different parameter sets to encode to current demand signal section, and aligns processed signal specific section and select to provide the IID parameter set of lowest overhead.Energy level between right front and rear sound channel is similar is possible, and energy level energy level similar but between right front and rear sound channel between front and rear L channel to differ widely be also possible.Provide delta coding and the coding of entropy afterwards of parameter, operation parameter q ₃and q ₄replace r ₃and r ₄to be more efficient.To having another signal segment of different characteristics, different parameter sets will provide lower bit-rate overhead.The characteristic of given signal segment, the present invention allows freely to switch between different parameters represents, to minimize the bit-rate overhead of present encoding signal segment.Between the different parameters of IID parameter, switch to obtain minimum ability that may bit-rate overhead, and what parameterized ability what provide that signaling arrangement represents current use being, is essential characteristic of the present invention.

In addition, the delta coding of parameter can complete in frequency direction or time orientation, and the delta coding between different parameters is also like this.According to the present invention, parameter can be carried out delta coding about other arbitrary parameter, if provide signaling arrangement to represent used specific delta coding.

The interesting feature of encoding scheme is the ability of carrying out scalable coding arbitrarily.The bit stream that this means coding can be divided into some different layers.Core layer can be by oneself decoding, and higher layer can decodedly strengthen the core layer signal of decoding.To different situations, the available number of plies may change, but as long as core layer can be used, demoder just can generate output sample.Above-mentioned use r ₁to r ₅the parametrization of the multi-channel encoder of parameter, makes them be very suitable for scalable coding.So, can in enhancement layer, store for example data of two surround channels (A and E) (being parameter r3 and r4), and store the parameter (by parameter r1 and r2 represented) corresponding with analog channel in core layer.

In Figure 10, summarize according to realization that can spread bit stream of the present invention.Layer bitstreams is by 1001 and 1002 diagrams, and wherein 1001 is core layers, has the mixed signal of waveform coding contracting and reappears the required parameter r of preposition sound channel (102,103 and 104) ₁and r ₂.1002 illustrated enhancement layers have the parameter for reappearing rearmounted sound channel (101 and 105).

Another importance of the present invention is the use of decorrelator in multichannel configuration.Use thought situation for one or two sound channel in PCT/SE02/01372 document of decorrelator to be described in detail.But in the time this theory being expanded to more than two sound channels, the problem that some the present invention solve has occurred.

Elementary mathematics shows, in order to obtain M the signal of decorrelation mutually from N signal, needs M-N decorrelator, and wherein all different decorrelators are all the functions that generates mutually orthogonal output signal according to public input signal.Typical decorrelator is all-pass or the wave filter that approaches all-pass, and given input x (t) just generates output y (t), wherein E[|y| ²]=E[|x| ²], and be almost 0 simple crosscorrelation E[yx ^*].The design of the decorrelator that perceptual standard in addition also enters, the example of some methods for designing can be will minimize the characteristic of comb filter in the time that original signal is added to de-correlated signals, and minimizes the impact of the long shock response of momentary signal sometimes.The artificial reverberator of decorrelator utilization (artificial reverberator) of some prior aries carrys out decorrelation.Prior art also comprises the mark time delay that the phase place by for example revising multiple sub-band samples realizes, so obtain more time diffusion to obtain higher echogenic density.

The method that the present invention proposes is that the decorrelator of modification based on reflection, realizes the multichannel decorrelator that generates the output signal of mutual decorrelation according to public input signal.Given identical input, if the output y of two decorrelators ₁and y (t) ₂(t) have and equal 0 or be almost 0 simple crosscorrelation, these two mutual decorrelations of decorrelator so.Suppose that input is stationary white noise, draw thus shock response h ₁and h ₂must be orthogonal, E[h ₁h ₂ ^*] be 0 or be almost 0.The decorrelator group of the mutual decorrelation of mode can be built by some modes in pairs.The efficient way of revising is the phase rotation coefficient q changing as a mark time delay part.

Prescribed phases twiddle factor of the present invention can be a part for lag line in all-pass filter, or is exactly whole fractional delay.In the later case, the method be not limited to all-pass or and the similar wave filter of reverberator, and can also be applied in the simple time delay that for example comprises mark time delay part.In decorrelator, all-pass filter link can be described as in Z territory:

$H\left(z\right)=\frac{{\mathrm{qz}}^{-m}-a}{1-{\mathrm{aqz}}^{-m}}$

Phase rotation coefficient that wherein q is complex values (| q|=1), m is the length of delay line in sample, and a is filter coefficient.Because stable reason, the amplitude of filter coefficient is restricted to | a| < 1.But, use the filter coefficient a '=-a of another selection, define new reverberator, there is identical reflection loss characteristic, significantly uncorrelated but output and uncorrected reverberator are exported.In addition, the correction of phase rotation coefficient q, can be by for example adding the phase bias q '=qe of constant ^jCcarry out.Constant C can be used as the phase bias of constant, or can carry out convergent-divergent, makes the constant time biasing of its all frequency bands that are applied to corresponding to it.Phase bias constant C can be also the random number not identical to all frequency bands.

According to the present invention, generate n sound channel from m sound channel, by being that to be applied to size be that the column vector of (m+p) × 1 signal is carried out to the upper mixed matrix H of n × (m+p) by size

$y=\left[\begin{array}{c}m\\ s\end{array}\right]$

Wherein m is the signal of m the mixed coding of contracting, and p signal decorrelation each other between two in s, and with m in all signal decorrelations.These de-correlated signals produce in the signal from m by decorrelator.N reconstruction signal a ', b ' ... be then comprised in column vector

In x '=Hy.

In Figure 11, illustrate above, wherein de-correlated signals is produced by decorrelator 1102,1103 and 1104.Upper mixed matrix H provides by 1101, and it acts on vectorial y and provides output signal x '.

Allow R=E[xx ^*] be the correlation matrix of original signal vector, allow R '=E[x ' x ' ^*] be the correlation matrix of reconstruction signal.Here and hereinafter, for matrix or the vectorial X with complex item, X ^*represent adjoint matrix, the complex conjugate transposition of X.

The diagonal line of R contain energy value A, B, C ... and according to energy quota defined above, be decoded to total energy level.Because R ^*=R, only have n (n-1)/2 different from diagonal line cross correlation value, these numerical value contain the information of carrying out all or part of reconstruction by adjusting upper mixed matrix H.The reconstruction of complete dependence structure is corresponding to the situation of R '=R.The reconstruction of correct energy level only corresponding R ' and R in the equal situation of diagonal line.

In the situation that obtains n sound channel from m=1 sound channel, the reconstruction of complete dependence structure can be by realizing with the decorrelator of p=n-1 mutual decorrelation.Upper mixed matrix H satisfies condition

$H{H}^{*}=\frac{1}{M}R$

Wherein M is the energy of single transmission signal.Because R is positive semidefinite, well-known such solution exists.In addition, the design of H has been left to the degree of freedom of n (n-1)/2, this is used for obtaining the upper mixed how useful characteristic of matrix in the present invention.Center standard is that H wants steadily the dependency degree of transmitted related data.

In parametrization, a kind of facilitated method of mixed matrix is H=UDV, and wherein U and V are orthogonal matrixes and D is diagonal matrix.Square can being chosen as with the eigenwert of R/M of the absolute value of D equates.Omit V also to eigenwert sequence, make maximum value be applied to the first coordinate, this will minimize the gross energy of de-correlated signals of output.Orthogonal matrix U is real matrix, carrys out parametrization by n (n-1)/2 anglec of rotation.With the form transmission related data of n the diagonal angle numerical value of these angles and D, the steady dependence of the H wanting will directly be provided.But because energy datum has to be converted into eigenwert, this mode has been sacrificed extensibility.

The second method that the present invention lectures comprises, by according to R=GR ₀g definition Normalized Correlation Moment battle array R ₀, in R, energy part is separated with relevant portion, wherein G is diagonal matrix, its diagonal line numerical value equals the square root of the diagonal entry of R, namely

and R ₀diagonal entry is all 1.Make H ₀for just handing in mixed matrix, mixed in preferred normalization that it equates at energy, the complete incoherent situation of signal has been given a definition.These examples that preferably above mix matrix have:

$\frac{1}{\sqrt{2}}\left[\begin{array}{cc}1& -1\\ 1& 1\end{array}\right],$ $\frac{1}{2}\left[\begin{array}{ccc}1& 1& \sqrt{2}\\ 1& 1& -\sqrt{2}\\ \sqrt{2}& -\sqrt{2}& 0\end{array}\right],$ $\frac{1}{2}\left[\begin{array}{cccc}1& 1& 1& 1\\ 1& 1& -1& -1\\ 1& -1& -1& 1\\ 1& -1& 1& -1\end{array}\right].$

Then upper mixed being defined as

wherein matrix S is SS ^*=R ₀solution.This solution is to R ₀the dependence of middle normalized crosscorrelation value is chosen to be continuous, makes at R ₀in the situation of=I, S equals unit matrix I.

N sound channel is divided into to contain the sound channel group of channel number be still less the facilitated method of rebuilding part simple crosscorrelation structure.According to the present invention, for the situation from 1 sound channel to 5.1 sound channels, particularly advantageous grouping is { a, e}, { c}, { b, d }, { f }, wherein to group c}, { f } do not apply decorrelation, and group a, e}, b, and d} by the contracting to identical mixed/decorrelation right on mix to produce.To these two subsystems, in complete incoherent situation, in preferred normalization, mix and elect as respectively:

$\frac{1}{\sqrt{2}}\left[\begin{array}{cc}1& -1\\ 1& 1\end{array}\right],$ $\frac{1}{\sqrt{2}}\left[\begin{array}{cc}1& 1\\ 1& -1\end{array}\right],$

So only have two to be transmitted and to rebuild in whole 15 simple crosscorrelation, namely sound channel { a, e} and { b, the simple crosscorrelation between d}.In the term of above-mentioned use, this is for n=6, the design example of m=1 and p=1 situation.Upper mixed matrix H size is 6 × 2, and wherein the 3rd row of secondary series and two elements of the 6th row are 0, corresponding to output c ' and f '.

The third method that what the present invention lectured be used for merging de-correlated signals is simpler viewpoint: each output channels has a different decorrelator, produces de-correlated signals s _a, s _b....Reconstruction signal is formed as

etc...

Parameter

control appears at output channels a ', b ' ... in the amount of de-correlated signals.With these, the form by angle is transmitted related data.Be easy to calculate, for example, normalized crosscorrelation between sound channel a ' and the b ' of generation equals product

because the number of paired simple crosscorrelation is n (n-1)/2 and has n decorrelator, if in general n > 3, can not mate with given dependency structure in this way, be a kind of very simple and stable coding/decoding method but advantage is it, and directly control the generation quantity that appears at the de-correlated signals in each output channels.This mixing that makes de-correlated signals is based on perceptual standard and combines the standard of the energy water adjustment that for example sound channel is right.

For the situation that obtains n sound channel from the sound channel of m > 1, correlation matrix R _y=E[yy ^*] can not be assumed to be again diagonal matrix, and by R '=HR _yh ^*while matching objective matrix R, also must consider this point.Because R _ythere is Block matrix structure, can be reduced to

$R_y=\left[\begin{array}{cc}{R}_m& 0\\ 0& R_s\end{array}\right],$

Wherein R _m=E[mm ^*] and R _s=E[ss ^*].In addition, suppose the decorrelator of mutual decorrelation, i.e. matrix R _sfor diagonal matrix.Note, this also affects the upper mixed design about correct energy Reconstruction.Solution is to calculate at demoder, or mixes the dependency structure R of signal from scrambler transmission and contracting _mrelevant information.

For the situation that obtains 5.1 sound channels from 2 sound channels, upper mixed preferred method is

$\left[\begin{array}{c}{a}^{\&prime;}\\ b^{\&prime;}\\ c^{\&prime;}\\ d^{\&prime;}\\ e^{\&prime;}\\ f^{\&prime;}\end{array}\right]=\left[\begin{array}{cccc}{\mathrm{<figure-callout\; id="11"\; label="h"\; filenames="HSA00000475998500061.png,HSA00000475998500082.png"\; state="\{\{state\}\}">h</figure-callout>}}_{11}& 0& {\mathrm{<figure-callout\; id="13"\; label="h"\; filenames="HSA00000475998500022.png,HSA00000475998500061.png"\; state="\{\{state\}\}">h</figure-callout>}}_{13}& 0\\ h_{21}& 0& h_{23}& 0\\ h_{31}& h_{32}& 0& 0\\ 0& h_{42}& 0& h_{44}\\ 0& h_{52}& 0& h_{54}\\ h_{61}& h_{62}& 0& 0\end{array}\right].\left[\begin{array}{c}{m}_1\\ m_2\\ S_1\\ {\mathrm{<figure-callout\; id="2"\; label="S"\; filenames="HSA00000475998500041.png"\; state="\{\{state\}\}">S</figure-callout>}}_2\end{array}\right],$

Wherein s ₁according to m ₁=l _ddecorrelation obtain, and s ₂according to m ₂=r _ddecorrelation in obtain.

Here be considered to right simple crosscorrelation, by group, { a, b} is with { d, e} is as independently 1 → 2 sound channel system processing.For sound channel c and f, adjust weight and make

E[|h ₃₁m ₁+h ₃₂m ₂| ²]＝C，

Ｅ[|h ₆₁m ₁+h ₆₂m ₂| ²]＝F.

The present invention is directed to storage or the transmission of different types of system, signal (analog or digital), can use codec arbitrarily to realize with the form of hardware chip and DSP.Fig. 2 and Fig. 3 have shown of the present invention may realization.In this example, shown the system being operated under six input signals (5.1 channel configuration) condition.What in Fig. 2, demoder one side showed is the analog input signal of each sound channel, and they are converted into digital signal 201, and to use bank of filters be that each sound channel analyzes 202.The output of bank of filters is supplied to around scrambler 203, and it comprises parameter generators, carries out one or two sound channel that the mixed generation of contracting audio coder 205 is encoded.In addition, be extracted out, and according to the present invention the temporal frequency grid of summary data and adopted which kind of parameterized control data to be extracted out 204 around for example IID of parameter and ICC parameter according to the present invention.Lecture as the present invention, the parameter extracting is encoded 206, or switches between different parametrizations, or with extendible mode configuration parameter.The mixed signal 208 of contracting around parameter 207, control signal and coding is re-used 209 to serial bit stream.

In Fig. 3, show a kind of typical demoder realization, be namely used to form the equipment that multichannel is rebuild.Here suppose the signal of audio decoder output frequency domain representation, the signal of for example exporting from the efficient AAC demoder of MPEG-4 before QMF synthesis filter banks.Serial bit stream is demultiplexed 301, being supplied to around data decoder 303 around data of coding, and the sound channel of contracting shuffling code is supplied to audio decoder 302, is the efficient AAC demoder of MPEG-4 in this example.To around data decode, and supplied with surround decoder device 305 around data decoder, it comprises mixed device, the contracting mixing sound road of upper mixed device based on decoding, linking number according to this and control signal reappear six sound channels.Frequency domain output from surround decoder device is synthesized 306 in time-domain signal, is then converted to simulating signal by digital to analog converter (DAC) 307.

Although the present invention is described about the generation of balance parameters and use, here require emphasis, with the sound channel for drawing balance parameters to the same grouping of dividing into groups, also can for calculate two sound channels between sound channel between relevant parameters or " width " parameter.In addition, mistiming or one between sound channel " phase place prompting ", also can apply and calculate identical sound channel to obtaining for balance parameters.In receiver one side, these parameters can be used as other parameter or the alternative parameter of balance parameters, form multichannel and rebuild.As selection, except between the definite sound channel of the reference sound channel by other amplitude difference, also can use between sound channel relevant parameters or or even sound channel between the mistiming.But, consider the characteristic of extensibility of the present invention, as the discussion of Figure 10 a and Figure 10 b, preferred method is that all parameters are used to identical sound channel pair, make in extendible bit stream, each extension layer comprises all parameters for rebuilding output channels subgroup, and these parameters are produced by corresponding extension layer, the extension layer of summarizing in row as second from the bottom in the table of Figure 10 b.When each sound channel between while only having correlation parameter or mistiming parameter are calculated and are transferred to demoder, the present invention is useful.In this case, in the time carrying out multichannel reconstruction, range parameter Already in demoder place uses.

Depend on the specific implementation requirement of method of the present invention, method of the present invention can realize with hardware or software.Implementation can be to use digital storage media, particularly stores disk or the CD of the control signal that can be read by electric mode, and storage medium cooperates with programmable computer system, and method of the present invention is carried out.In general, the present invention can be also computer program, and program code is stored on machine-readable carrier, in the time that computer program moves on computers, can carry out method of the present invention.In other words, method of the present invention is computer program, and this program contains the program code of carrying out the method for at least a kind of invention while operation on computers.

Claims (4)

1. the equipment for generation of range parameter, described range parameter is the additional parameter having outside the Parametric Representation of multi-channel signal of multiple original channel, described Parametric Representation comprises parameter set, described parameter set is realized multichannel and is rebuild in the time using together with at least one contracting mixing sound road, and described equipment is configured to calculate range parameter (r _m), described range parameter be energy in main contracting mixed a time samples piece and a frequency band with Parametric Representation institute based on parameter contracting the time samples piece and the frequency band that mix in the ratio of energy,

Wherein, described range parameter and described parameter set and main contracting are mixed can transfer to receiver.

2. a Parametric Representation for using with parameter set produces and has the equipment that the re-establishing multiple acoustic track of original multi-channel signal of at least three original channel represents, described parameter set is realized multichannel and is rebuild in the time using together with at least one contracting mixing sound road, wherein, except Parametric Representation, provide in addition range parameter as additional parameter, described range parameter be energy in main contracting mixed a time samples piece and a frequency band with Parametric Representation institute based on parameter contracting the time samples piece and the frequency band that mix in the ratio of energy, described equipment comprises:

Amplitude modifier (902), for by single basic sound channel or multiple basic sound channel being weighted with range parameter, is used range parameter to described single basic sound channel or the correction of multiple basic sound channel applies amplitude.

3. the method for generation of range parameter, described range parameter is the additional parameter having outside the Parametric Representation of multi-channel signal of multiple original channel, described Parametric Representation comprises parameter set, described parameter set is realized multichannel and is rebuild in the time using together with at least one contracting mixing sound road, and described method comprises:

Calculate range parameter (r _m), described range parameter be energy in main contracting mixed a time samples piece and a frequency band with Parametric Representation institute based on parameter contracting the time samples piece and the frequency band that mix in the ratio of energy; And

Generation comprises the output data that described range parameter and described parameter set and main contracting are mixed.

4. a Parametric Representation for using with parameter set produces and has the method that the re-establishing multiple acoustic track of original multi-channel signal of at least three original channel represents, described parameter set is realized multichannel and is rebuild in the time using together with at least one contracting mixing sound road, wherein, except Parametric Representation, provide in addition range parameter as additional parameter, described range parameter be energy in main contracting mixed a time samples piece and a frequency band with Parametric Representation institute based on parameter contracting the time samples piece and the frequency band that mix in the ratio of energy, described method comprises:

By single basic sound channel or multiple basic sound channel being weighted with range parameter, use range parameter to described single basic sound channel or the correction of multiple basic sound channel applies amplitude, carry out the upper mixed correction multichannel reconstruction obtaining thereby obtained by the concentrated parameter of operation parameter.

Images